282 
Fishery Bulletin 11 7(4) 
Table 2 
All canonical correspondence analysis (CCA) axis loadings (CCA1 and CCA2) of diet items and constraining variables 
used to identify patterns in the diet of blue catfish (Ictalurus furcatus) collected during 2013-2016 in 4 tributaries 
to Chesapeake Bay in Virginia—the James, Pamunkey, Mattaponi, and Rappahannock Rivers. Variables include 
salinity zone, season (spring and summer), and predator total length. Separate models were constructed for each river. 
James Pamunkey Mattaponi Rappahannock 
Variable 
CCA1 
CCA2 
CCA1 
CCA2 
CCA1 
CCA2 
CCA1 
CCA2 
Salinity zone 
-0.624 
0.303 
0.816 
-0.312 
0.150 
0.013 
0.284 
0.105 
Spring 
0.720 
0.574 
-0.482 
0.759 
-0.415 
-0.868 
-0.665 
0.702 
Summer 
-0.036 
-0.756 
-0.420 
-0.870 
-0.092 
0.898 
0.029 
-0.973 
Total length 
-0.050 
0.530 
0.286 
-0.289 
-0.194 
0.391 
-0.041 
-0.078 
(P<0.001), and all predictors were significant (P<0.05), 
with the exception of salinity in the model for predation on 
American eel (Table 3). All GAMs had acceptable predic¬ 
tive performance, with areas under the ROC of 0.84-0.86 
(Table 2; Pearce and Ferrier, 2000). 
Our models demonstrate that depleted alosines, Amer¬ 
ican shad and river herring, were most susceptible to 
predation by blue catfish in tidal freshwater areas. As 
many as 4% of stomachs from blue catfish were expected 
to contain these taxa in certain areas (e.g., tidal freshwa¬ 
ter stretches of the James River; Fig. 3). Our model also 
revealed that large blue catfish consumed more alosines, 
and as many as 8% of stomachs from 700-1000-mm-TL 
blue catfish were predicted to contain American shad or 
river herring in the James River. Seasonally, the probabil¬ 
ity of predation upon depleted Alosa species was greatest 
in April, with another increase in predation during Octo¬ 
ber. Overall, predicted predation on alosines was highest 
in the James and Rappahannock Rivers. In the imperiled 
alosine GAM, river herring were the most commonly con¬ 
sumed species group in both rivers, although American 
shad were found in more stomachs of blue catfish from the 
Rappahannock River than in those from the James River 
(Schmitt et al., 2019). 
Our model suggests that predation on blue crab 
increases at higher salinities. Nearly 30% of stomachs 
from blue catfish were predicted to contain blue crab in 
S P levels >8 in the James River, and predicted percent 
occurrence of blue crab in stomachs was typically less than 
5% in the other rivers. Large blue catfish consumed blue 
crab more frequently, and model predictions indicate that 
catfish between 600 and 900 mm TL were most likely to 
consume blue crab (Fig. 4). Model predictions also indicate 
that predation on blue crab was greatest during the late 
summer and into fall (August-October). 
Predation on American eel was uncommon, and pre¬ 
dicted occurrence in stomachs of blue catfish was <5% in 
all modeled scenarios (Fig. 5). Predation on American eel 
was not significantly correlated with salinity (P>0.05), 
although it was correlated with predator TL and month 
(P<0.02). Model predictions indicate that medium and 
large blue catfish (500-900 mm TL) were the most likely 
to consume American eel. Seasonally, predicted occurrence 
was highest during spring and fall, particularly in April 
and October (Fig. 5). 
Discussion 
In all rivers, the diet of blue catfish varied with season , 
salinity, or both. These 2 factors also influence the struc¬ 
ture of assemblages of organisms in Chesapeake Bay 
(Wagner and Austin, 1999; Jung and Houde, 2003; Lippson 
and Lippson, 2006). These relationships are intuitive 
because species assemblages vary drastically along the 
salinity gradient and some species are only available sea¬ 
sonally (Wagner and Austin, 1999; Jung and Houde, 2003; 
King et al., 2005). For example, aquatic macrophytes, 
which are commonly found in stomachs of blue catfish 
(Schmitt et al., 2019), are generally only available during 
the warmer months (Moore et al., 2000). Other potential 
prey in tributaries of Chesapeake Bay include adult Amer¬ 
ican shad, hickory shad (A. mediocris), and river herring 
that make upstream spawning migrations during spring 
(Garman and Nielsen, 1992; Schmitt et al., 2017). Addi¬ 
tionally, juveniles of these taxa emigrate from these rivers 
during the late summer and autumn months (Hoffman 
et al., 2008). Blue crab and American eel also make sea¬ 
sonal movements in these rivers (Wenner and Musick, 
1974; Aguilar et al., 2005), and our models revealed 
increased predation during these migratory periods. 
Lastly, although interannual variation is likely an import¬ 
ant driver of dietary patterns for blue catfish, stomach 
contents were pooled across years for each month to 
increase sample sizes for our study. Other diet studies 
have reported strong interannual trends that mirror fluc¬ 
tuations in prey abundance (Latour et al., 2008). 
Multivariate analyses identified consistent, length- 
related shifts from omnivory to piscivory, a finding con¬ 
sistent with previous work (Schmitt et al., 2017, 2019). In 
general, small blue catfish feed primarily on macro¬ 
phytes and benthic invertebrates, and large blue catfish 
become more piscivorous (see Schmitt et al., 2019). The 
size at which this shift to piscivory occurs varies from 
